Deep Learning Limits
_https://www.wired.com/story/greedy-brittle-opaque-and-shallow-the-downsides-to-deep-learning
According to skeptics like Marcus, deep learning is **greedy, brittle, opaque, and shallow. **
The systems are greedy because they demand huge sets of training data.
- BrainBlocks doesn’t require much data
- synthetic data
- recognition focused and not classification
Brittle because when a neural net is given a “transfer test”—confronted with scenarios that differ from the examples used in training—it cannot contextualize the situation and frequently breaks.
- Open Set Recognition
They are opaque because, unlike traditional programs with their formal, debuggable code, the parameters of neural networks can only be interpreted in terms of their weights within a mathematical geography. Consequently, they are black boxes, whose outputs cannot be explained, raising doubts about their reliability and biases.
- Human Explainability
Finally, they are** shallow** because they are programmed with little innate knowledge and possess no common sense about the world or human psychology.
- Leverage Knowledge or Other Models
Gary Marcus
Marcus, Gary. “Deep learning: A critical appraisal.” arXiv preprint arXiv:1801.00631 (2018).
_https://medium.com/@GaryMarcus/in-defense-of-skepticism-about-deep-learning-6e8bfd5ae0f1
Innateness, AlphaZero, and Artificial Intelligence_ _ https://arxiv.org/pdf/1801.05667
_ _ _ Deep Learning: A Critical Appraisal_ _ _https://arxiv.org/abs/1801.00631